This study investigated the effects of using three strategies for reading on the comprehension of fiction and nonfiction selections by fifth graders. The subjects (N=270) were grouped into three different reading ability levels and randomly assigned to the three strategies for reading: (1) subjects read the entire selection and then answered 10 questions (RT), (2) subjects read 10 questions, then read the selection and answered 10 questions (QRT), and (3) subjects were given the 10 questions, each with four multiple choice answers; they were asked to select an answer for each question, read the selection, and again answer 10 questions (QSRT). Data were analyzed by Analysis of Variance, Duncan's Multiple Range Statistic, t-tests and Chi square. Findings were that (1) neither of the three strategies is more effective than the other, but a trend favored the QSRT strategy for the above average and average readers, (2) the reading strategies did not make a difference of statistical significance when performance on the fiction and nonfiction selections was compared, (3) regardless of reading ability level or the strategy used, the subjects performed better when reading fiction rather than nonfiction, and (4) the subjects achieved higher scores on the pretest than on the post test. (Author/LK)